TWI543634B - Method and computer program product of processing voice segment and hearing aid - Google Patents

Description

Method for processing sound segments and computer program products and hearing aids

The present invention relates to a method and a hearing aid for improving the correctness of a hearing impaired person.

Since most hearing impaired persons (including elderly people) have problems with high-frequency sounds, hearing aids basically perform low-frequency processing of high-frequency sounds.

Prior art U.S. Patent No. 6,577,739 "Apparatus and methods for proportional audio compression and frequency shifting" compresses an audio signal to a hearing loss for a specific range, and the technique is to compress the entire sound. Will make the sound distortion seriously.

Prior art U.S. Patent No. 7,609,841 "Frequency shifter for use in adaptive feedback cancellers for hearing aids" improves the frequency shifting method of the prior art by shifting only the high frequency portion of the signal (up or down) and shifting the frequency shift ratio Limited to 6%.

Prior art U.S. Patent No. 7,580,536, "Sound Enhancement for hearing-impaired listeners" refers to frequency shifting of high frequency portions of a signal.

Applicant's US Patent No. US20120140964 "Methodand hearing aid for enhancing the accuracy of sounds heard by ahearing-impaired listener", which discloses which high frequency parts should be subjected to low frequency processing.

When the input voice is processed in real time, the input voice is first cut into sound segments, and then it is judged whether or not the sound segments are to be subjected to low frequency processing. A condition often encountered, such as a vowel sound, is followed by a high frequency consonant (eg, "at"). Traditionally, it belongs to "a". All sound segments are completely unprocessed, and all sound segments belonging to "t" are treated with low frequency processing. However, between "a" and "t" are often non-high frequency or non- Low-frequency sound, but once the hearing aid determines that it is high-frequency, it performs low-frequency processing. If it is judged to be low-frequency, it does not perform low-frequency processing. However, such intermediate mixed sound is sometimes difficult to be classified, resulting in "at" processing. In the process, there is sometimes a smooth sound between "a" and "t", and the above-mentioned predecessor does not discuss how to deal with this problem.

The primary object of the present invention is to provide a method of processing a mixed sound segment that is a segment of sound between a high frequency and a low frequency sound.

To achieve the above objects, the present invention is applied to a hearing aid, and the method includes:

Checking the type of a sound segment, wherein the sound segment is divided into at least three types: a low frequency type, a mixed type, and a high frequency type, wherein the high frequency type has the following characteristics: the energy above A Hz is greater than X% and Less than 100%, of which 1200≦A≦3000, 50≦X≦60. The hybrid type has the following characteristics: the energy above A Hz is less than X% and greater than Y%, where X-30 ≦ Y ≦ X-5. The low frequency type has the following characteristics: the energy above A Hz is greater than or equal to 0% and less than Y%.

When the sound segment is of the high frequency type, a first low frequency processing is performed on the sound segment, the first low frequency processing is at least for a partial sound segment of the sound segment above B Hz, and the sound segment is at B The energy of the partial sound segment above Hertz does not reduce the energy before performing the first low frequency processing, wherein 2000 ≦ B ≦ 5000;

When the sound segment is of the hybrid type, a second low-frequency processing is performed on the sound segment, and the second low-frequency processing is at least for a partial sound segment of the sound segment above B Hz, and the sound segment is at B Hz The energy of the above partial sound segments will reduce the energy before the completion of the second low frequency processing. According to an embodiment of the present invention, when the sound segment is of the hybrid type, when the energy ratio above AHz is smaller, the energy of the sound segment is lower than the energy of the partial sound segment above BHz at the second low frequency processing. The more the percentage.

In order to enable the reviewing committee to better understand the technical contents of the present invention, the preferred embodiments are described below.

Please refer to FIG. 1 for the overall architecture diagram of the hearing aid of the present invention.

The hearing aid 10 of the present invention includes a sound receiver 11, a sound processing module 12, and a speaker 13. The sound receiver 11 is configured to receive the input sound 20, which is processed by the sound processing module 12 and then broadcast by the speaker 13 to the hearing impaired person 81. The sound receiver 11 can be any sound-receiving device such as a microphone, and the speaker 13 (which can also have an amplifier) can be any sound-transmitting device such as a headphone, but the present invention is not limited to the devices listed above. The sound processing module 12 is generally composed of a sound processing chip matching control circuit and an amplifying circuit; and can also be a solution technology composed of a processor, a memory matching control circuit, and an amplifying circuit. The focus of the sound processing module 12 is to amplify the sound signal, filter the noise, change the sound frequency composition, and the processing required to achieve the object of the present invention. Since the sound processing module 12 can be matched with a new firmware or a new firmware or Software, so the hardware architecture of the sound processing module 12 will not be described again. The so-called hearing aid 10 of the present invention can be, for example, a special machine for tailoring hardware, or a small computer such as a PDA, a mobile phone, a hearing aid earphone (such as a Bluetooth earphone, a chip or processor for processing audio), or a smart phone. And the personal computer implements the present invention in a software program.

Next, please refer to FIG. 2, which is a flow chart of steps of the sound processing module of the present invention. Please also refer to FIG. 3 to FIG. 9 together to understand the present invention.

Step 201: Receiving the input sound 20, this step is completed by the sounder 11.

Step 202: Cut the input sound 20 into a plurality of sound segments 21. The length of each sound segment is recommended to be between 0.0001 and 0.1 seconds. In the experiment with I-Phone4 as a hearing aid (the software produced according to the present invention is executed in I-Phone4), the length of the sound segment is about 0.0001. The effect is good between ~0.1 seconds.

Step 203: Check to check if a sound segment is of a high frequency type?

Please refer to Fig. 4, the input sound 20 shows the sound segments of different patterns, and the darker part in Fig. 4 indicates that the more energy, for example, the high frequency area A3 is more energy at 2500 Hz (which is different from different languages). The high frequency area A3 is usually a high frequency consonant (such as the pronunciation of T, S in English) in speech, such as a high frequency type sound segment 21c. The low frequency area A1 has more energy below 2500 Hz, and the low frequency area A1 is usually the vowel (such as the English O, A pronunciation) and the low frequency sub sound (such as the English Z, M pronunciation) in the voice, such as the low frequency type sound. Segment 21a. In the junction area A2 (low-frequency sound-to-high-frequency sound, or high-frequency sound-to-low-frequency sound), in this embodiment, the intersection area of the vowel and the high-frequency consonant A2 often has a condition of mixing high-frequency consonants and vowels, such as a mixed type sound. Segment 21b.

The high frequency type sound segment 21c has the following characteristics: the energy above A Hz is greater than X% and less than 100%, of which 1200 ≦A ≦ 3000, 50 ≦ X ≦ 60. The A value and the X value are ranged. The main reason is that if A is set relatively small, X will be larger, and the accuracy of the experimental results will not be too different in some ranges. In addition, the A value and the X value and language. The type is also related, and it is also related to the pronunciation of male and female voices. In the Chinese experiment, one of the software programs developed by the inventor A=1500, X=55.

If YES in step 203, then step 204 is performed, and if no, step 205 is performed.

Step 204: Perform a first low frequency processing D1 on the sound segment. The first low-frequency processing D1 is at least for a partial sound segment of the sound segment above B Hz, and the energy of the partial sound segment of the sound segment above B Hz does not reduce energy before performing the first low-frequency processing D1, wherein 2000 ≦B≦5000.

The low frequency processing of the high frequency type sound segment is a conventional technique, and the low frequency processing basically uses voltage frequency or/and frequency shifting, for example, shifting the main energy portion of the relatively high frequency in the sound segment to the low frequency by frequency shifting. Or the main energy portion of the higher frequency in the sound segment is pressed to the low frequency by the frequency of the frequency, and of course, the frequency and the frequency shift are often performed together. For example, if the high frequency type sound segment 21c is mainly in the high frequency concentrated region 211c, 4000 Hz to 6000 Hz, the high frequency concentrated region 211c is compressed to 4000 Hz to 5000 Hz, and then shifted to 3500. Hertz ~ 4500 Hz (500 Hz down), that is, the high frequency concentration area 211c (4000 Hz ~ 6000 Hz) becomes the sound section 21c' of the concentrated area 211c' (3500 Hz ~ 4500 Hz). It is to be noted that the frequency and the frequency shift are conventional techniques. The above is only an example. For example, reference is made to the applicant's US Patent No. US20120140964, because the present invention is not a technique for improving the frequency and frequency shift, such as a frequency-frequency ratio. The range selected by the voltage frequency, the interval of the frequency shift, etc., will not be described in this case. In addition, the value of B depends on the condition of the hearing impaired person. The more severe the hearing impaired person, the smaller the value of B; and the slightly hearing impaired person can hear a higher frequency sound, so the value of B is larger.

It should be noted in step 204 that the energy of the partial sound segment of the sound segment above B Hz does not decrease the energy before the first low-frequency processing D1 is performed, that is, in the present embodiment, the high-frequency concentration region 211c performs the first low frequency. No energy was reduced before D1 was processed.

Step 205: Check if a sound segment is a mixed type? The hybrid type has the following characteristics: the energy above A Hz is less than X% and greater than Y%, where X-30 ≦ Y ≦ X-5. If YES in step 205, proceed to step 206; if not, if the sound segment is of a low frequency type (such as a vowel or a low frequency consonant), no processing is performed, and when the sound segment is of a low frequency type, the following characteristics are obtained: above A Hz The energy is greater than or equal to (≧) 0% and less than Y%.

Step 206: Perform a second low frequency processing D2 on the sound segment. That is, when the sound segment is the mixed type sound segment 21b, a second low frequency processing D2 is performed on the sound segment, and the second low frequency processing D2 is at least for a partial sound segment of the sound segment above B Hz, and the sound segment is The energy of the partial sound segment above B Hz reduces the energy before the completion of the second low frequency processing D2. The second low-frequency processing D2 is also to perform frequency-comprising or/and frequency-shifting for the high-frequency concentration area 211b, but the emphasis is on actually reducing the energy for the high-frequency concentration area 211b, and finally forming the frequency-reduced concentrated area 211b'. For example, the high-frequency concentration area 211b is first reduced in energy, then frequency-shifted or frequency-converted, and finally the frequency-reduced concentrated area 211b' is formed; or the high-frequency concentrated area 211b is firstly frequency-reduced, but the energy is reduced before the frequency shifting, Finally, the frequency shift is formed to form a down-converted concentration area 211b'. The above steps 202 to 206 are performed by the sound processing module 12.

The energy reduction can be the last energy reduction coefficient, and the energy reduction coefficient is less than 1 and greater than 0. Please refer to Figure 7. The smaller the energy above A Hz is, the more the energy reduction energy percentage of the partial sound segment above the B Hz is at the second low frequency processing. For example, if the mixed type sound segment is more than 40% (Y%) but less than 50% (X%), if the sound of more than 1500 Hz has a sound segment of 49%, the energy reduction coefficient is 0.9. That is, the high frequency concentration region 211b retains 90% of the energy for the frequency/shift. However, if the energy of more than 1500 Hz of a sound segment accounts for 47%, the energy reduction coefficient is 0.7, that is, the high frequency concentration region 211b retains 70% of the energy for the frequency/shift. However, if the energy of more than 1500 Hz of a sound segment accounts for 41%, the energy reduction coefficient is 0.1, that is, the high frequency concentration region 211b retains 10% of energy for frequency/shift. The energy reduction coefficient is linear in the above example (solid line in Fig. 7), but may also be nonlinear (dashed line in Fig. 7).

The significance of the second low-frequency processing is that the closer the mixed-type sound segment is to the high-frequency type sound segment, the more the mixed-type sound segment processing is handled like the high-frequency type sound segment, that is, there are more high-frequency regions. The energy is retained and moved to the low frequency region. When the mixed type sound segment is closer to the low frequency type sound segment, the mixed type sound segment processing method is more like the low frequency type sound segment processing mode, that is, the energy in the high frequency region is less moved to the low frequency region, that is, The low frequency area is relatively unchanged.

It should be noted that the above-mentioned processed sound segment is usually amplified by the speaker 13 (headphone, speaker, sound amplifier) so that the hearing impaired 81 can hear the sound. The above steps are described in the entire sound segment. The processing before the volume is amplified.

It should be noted that the above is only an embodiment, and is not limited to the embodiment. For example, those who do not depart from the basic structure of the present invention should be bound by the scope of the patent, and the scope of the patent application shall prevail.

10‧‧‧Hearing aids
11‧‧‧Audio
12‧‧‧Sound Processing Module
13‧‧‧Speakers
20‧‧‧ Input sound
21, 21a, 21b, 21c‧ ‧ sound segment
211c, 211b‧‧‧ high frequency concentration area
211c', 211b'‧‧‧ after frequency reduction
81‧‧‧The hearing impaired
D1‧‧‧First low frequency processing
D2‧‧‧second low frequency processing
A1‧‧‧Low frequency zone
A2‧‧‧ junction area
A3‧‧‧High frequency zone

1 is an overall architectural diagram of a hearing aid of the present invention. 2 is a flow chart showing the steps of the sound processing module of the present invention. Figure 3 is a schematic diagram of the input sound cut into a plurality of sound segments. Figure 4 shows the sound segments of different patterns. Figure 5 is a schematic diagram of processing a high frequency type sound segment. Figure 6 is a schematic diagram of processing a mixed type sound segment. Figure 7 is a schematic diagram of the energy reduction coefficient.

Step 201 to step 206

Claims (5)

A method for processing a sound segment is applied to a hearing aid, the method comprising: checking a type of a sound segment, wherein the sound segment is divided into at least three types: a low frequency type, a mixed type, and a high frequency type, wherein : The high frequency type has the following characteristics: the energy above A Hz is greater than X% and less than 100%, wherein 1200 ≦A ≦ 3000, 50 ≦ X ≦ 60; the hybrid type has the following characteristics: the energy above A Hz is less than X% and greater than Y%, where X-30≦Y≦X-5; the low frequency type has the following characteristics: the energy above A Hertz is greater than or equal to 0% and less than Y%; when the sound segment is of the high frequency type And performing a first low-frequency processing on the sound segment, wherein the first low-frequency processing is performed on at least a partial sound segment of the sound segment above B Hz, and the energy of the partial sound segment of the sound segment above B Hz is performed. There is no energy reduction before the first low frequency processing, wherein 2000 ≦B ≦ 5000; and when the sound segment is the hybrid type, a second low frequency processing is performed on the sound segment, and the second low frequency processing is at least The sound segment is in B a portion of the sound segment above, and the energy of the partial sound segment of the sound segment above B Hz reduces the energy before the completion of the second low frequency processing, and the energy of the sound segment is reduced in a portion of the sound segment above B Hz The energy percentage is less than 100% and greater than 0%. The method for processing a sound segment as described in claim 1, wherein when the sound segment is of the mixed type, the ratio of energy above A Hz is smaller. At the time of the second low-frequency processing, the energy reduction percentage of the partial sound segment of the sound segment above B Hz is more. A hearing aid for receiving an input sound and modifying the input sound for output to a hearing impaired person, the hearing aid comprising: a sound receiver for receiving the input sound; a sound processing module coupled to the sound The device is electrically connected, and the input sound is cut into a plurality of sound segments, and the types of the sound segments are checked, wherein each of the sound segments is divided into at least three types: a low frequency type, a mixed type, and a high frequency type. Wherein: the high frequency type has the following characteristics: the energy above A Hz is greater than X% and less than 100%, wherein 1200 ≦A ≦ 3000, 50 ≦ X ≦ 60; the hybrid type has the following characteristics: above A Hertz The energy is less than X% and greater than Y%, wherein X-30≦Y≦X-5; the low frequency type has the following characteristics: the energy above A Hertz is greater than or equal to 0% and less than Y%; when the sound segment is the high frequency In the case of the type, the sound processing module performs a first low frequency processing on the sound segment, the first low frequency processing is at least for a partial sound segment of the sound segment above B Hz, and the sound segment is above B Hz Part of the sound segment No reducing energy before performing the first low-frequency processing, wherein 2000≦B≦5000; and when the sound segment is the hybrid type, the sound processing module performs a second low-frequency processing on the sound segment, The second low-frequency processing is for at least a partial sound segment of the sound segment above B Hz, and the energy of the partial sound segment of the sound segment above B Hz reduces the energy before the second low-frequency processing is completed, and the sound is reduced. segment The percentage of energy reduction energy of a portion of the sound segment above B Hz is less than 100% and greater than 0%. According to the hearing aid of claim 3, when the sound segment is of the hybrid type, the smaller the energy ratio above AHz, the sound segment is above B Hz when the second low frequency processing is performed. The percentage of energy reduction in the partial sound segment is greater. A computer program product, when a hearing aid is loaded into the computer program and executed, the method of any one of claims 1 to 2 can be completed.